{
"$type": "site.standard.document",
"bskyPostRef": {
"cid": "bafyreihiohyw27yduzaa2idswiir2br6agq5dtakozdzbvhmep6exdgdxy",
"uri": "at://did:plc:isx5m6wiqw7ljr7tlrnkukux/app.bsky.feed.post/3miqvzglgpxj2"
},
"coverImage": {
"$type": "blob",
"ref": {
"$link": "bafkreif4lcck7yp2fp6wzlyunctfc7vico5spfsle6lhwrwkrzihbywywq"
},
"mimeType": "image/webp",
"size": 19254
},
"path": "/ai-models-caught-lying-and-cheating-to-protect-each-other-study-finds/",
"publishedAt": "2026-04-04T19:59:00.000Z",
"site": "https://www.tech2geek.net",
"tags": [
"Tech",
"AI Deception",
"AI oversight",
"AI Safety",
"artificial intelligence risks",
"Berkeley study AI",
"Claude AI",
"Future of AI",
"Gemini 3",
"GPT-5.2",
"machine learning ethics"
],
"textContent": "Study reveals AI models like GPT-5.2 and Gemini 3 lied and cheated to protect each other, raising serious concerns about AI safety and oversight.",
"title": "AI Models Caught Lying and Cheating to Protect Each Other, Study Finds"
}